LIVE MARKETS-AI spending spree by U.S. hyperscalers may obscure rising financial risks
Apollo Global Management Inc APO | 0.00 | |
Alphabet Inc. Class A GOOGL | 0.00 | |
Meta Platforms META | 0.00 | |
Blue Owl Capital OWL | 0.00 | |
Oracle Corporation ORCL | 0.00 |
Welcome to the home for real-time coverage of markets brought to you by Reuters reporters. You can share your thoughts with us at markets.research@thomsonreuters.com
AI SPENDING SPREE BY U.S. HYPERSCALERS MAY OBSCURE RISING FINANCIAL RISKS
Technology companies’ heavy spending to gain an edge in artificial intelligence could pose a systemic financial risk because of the aggressive structures used to obscure their true outlays and revenues, Mark Malek, chief investment officer at Siebert Financial, said.
The five largest U.S. hyperscalers — Alphabet GOOGL.O, Amazon AMZN.O, Meta META.O, Microsoft MSFT.O and Oracle ORCL.N — are set to spend more than $1 trillion combined on AI infrastructure in 2025 and 2026, according to a Bank for International Settlements (BIS) report.
However, that figure understates total spending because hyperscalers often do not directly build data centers or hold related debt on their balance sheets.
Instead, companies set up special purpose vehicles (SPVs) that secure financing from private credit firms such as Blackstone BX.N, Apollo APO.N, Blue Owl OWL.N and PIMCO. The hyperscalers then lease the completed facilities.
This structure allows companies to classify the costs as operating leases while keeping the associated debt off their balance sheets. Lease-related commitments not recorded as liabilities total about $662 billion across the five companies, Malek said, citing a Moody’s analysis.
"To put that in perspective, that’s roughly 113% of those same five companies’ most recent adjusted debt, meaning the stuff hiding off the books is actually bigger than the stuff sitting on it," he says.
Malek also flagged “circular financing,” a practice highlighted by the BIS in which chipmakers and hyperscalers take equity stakes in AI labs.
Those labs, in turn, commit to multi-year purchases of chips and computing capacity, often from the same companies that invested in them.
"That’s not organic demand pulling the AI buildout forward. That’s manufactured demand, and manufactured demand has a way of flattering every metric right up until the moment it doesn’t," he says.
While these practices do not pose an immediate threat to financial stability, investors should closely monitor hyperscalers’ free cash flow relative to capital expenditure guidance on earnings calls, he added.
"The moment that guidance gets trimmed, the entire chain underneath it reprices at once, and you want to see that coming rather than reading about it after the fact," he says.
(Chibuike Oguh)
*****
EARLIER ON LIVE MARKETS:
LEADERSHIP WIDENS AS SMIDS OUTPACE THE S&P 500 CLICK HERE
THE TIDE IS CHURNING: JOLTS, CONSUMER CONFIDENCE, HOME PRICES, CHICAGO PMI CLICK HERE
BOFA SAYS 'OLD-ECONOMY' CYCLICALS ARE THE BETTER BET NOW CLICK HERE
US STOCKS STEADY INTO QUARTER-END; SMALL CAPS EYE BEST FIRST HALF SINCE 1991 CLICK HERE
S&P 500 REBOUNDS INTO QUARTER-END AS BULLS EYE STRONG GAINS CLICK HERE
THERMOMETERS ARE THE NEW ECONOMIC FORECASTING TOOL CLICK HERE
GOLDMAN SAYS DON'T ROTATE FROM ASIA'S TECH WINNERS CLICK HERE
GOLD SLUMP MAY HAVE GONE TOO FAR FOR TRADERS CLICK HERE
STOXX NEAR PEAKS AS AI TRADE ROLLS ON CLICK HERE
BEFORE THE BELL: EUROPE EYES BEST QUARTER SINCE 2022 CLICK HERE
INVESTORS GO SHOPPING FOR Q3 CLICK HERE
